Amazon Ads can look deceptively simple from the outside: pick some keywords, set a budget, and wait for sales. In practice, a healthy account is closer to a living system dozens of small decisions about targeting, bidding, and structure, adjusted continuously based on what the data shows.
Spending more on ads doesn't automatically produce more profitable sales. Two accounts with identical budgets can produce very different results depending on how campaigns are structured, which search terms are being paid for, and how closely performance is monitored. This article walks through how our team approaches Amazon advertising day to day from campaign structure and targeting through to how we read the numbers using a real month of Sponsored Ads data from one of our managed accounts as a working example.
All figures in this article are taken directly from an Amazon Ads Campaign Manager dashboard for one account we manage. Client and account details have been removed. No metrics have been invented or estimated.
Our Amazon Ads Approach
Every account we manage moves through the same core cycle, whether it's a new product launch or an established brand with years of sales history:
- Research - understanding the product, its competitors, and how customers actually search for it.
- Campaign setup - building a structure that separates discovery from control.
- Targeting - choosing keywords, products, and audiences deliberately, not broadly.
- Monitoring - reviewing search-term and performance reports on a regular schedule.
- Optimization - adjusting bids, budgets, and targets based on what the data shows.
- Scaling - putting more budget behind what's already working, rather than spreading spend thinly.
Each of these steps depends on the ones before it. Skipping research means campaigns launch with guesses instead of evidence; skipping monitoring means budget keeps flowing to underperforming targets long after the data has flagged a problem.
The Four Ad Formats We Work With
Amazon groups its advertising tools into a few core formats. We typically use a combination of these, rather than relying on just one, because each plays a different role in getting a shopper from "never seen this product" to "bought it."
Sponsored Products
Sponsored Products are the ads that appear directly in Amazon search results and on product detail pages, styled to look like regular listings but marked as "Sponsored." They're usually the foundation of an account because they put a specific product directly in front of a shopper who is already searching with buying intent.
We run these using a mix of Auto and Manual campaigns (explained below), which lets us combine Amazon's own discovery with the control we get from hand-picked targeting.
Sponsored Brands
Sponsored Brands appear as a banner-style ad above or within search results, typically showing a brand logo, a headline, and a small selection of products. They're useful for building brand recognition rather than pushing a single product, and they work well once we already have a sense of which keywords convert, since Sponsored Brands campaigns are usually built around proven search terms and, where relevant, audience targeting.
Sponsored Display
Sponsored Display ads can appear both on and off Amazon, and are commonly used for retargeting reaching shoppers who viewed a product (or a similar one) but didn't purchase. We also use product and audience targeting here to keep a brand visible to people who are already in a buying mindset for that category. Sponsored Display complements Sponsored Products by re-engaging shoppers rather than competing with them for the same search-result placements.
Amazon Video Ads
Video ads capture attention in a way static images can't, which makes them well suited to the earlier, awareness-building stage of the customer journey. We treat video as a funnel-top tool: it introduces a product to a wider audience, while Sponsored Products and Sponsored Brands do more of the work at the consideration and conversion stages.
Where each format sits in the customer journey
Auto vs. Manual vs. Self-ASIN Campaigns
Within Sponsored Products, we lean on three campaign types, each with a different job.
Auto campaigns hand targeting decisions to Amazon. Amazon looks at the product's title, category, and attributes, and automatically matches it to relevant search terms and other products. We use Auto campaigns mainly as a research tool they surface search terms and targeting opportunities we might not have thought to bid on manually, especially for newer listings with limited sales history.
Manual campaigns are where we take the keywords and targets that research (and Auto campaign data) has proven relevant, and place them into a controlled campaign where we set the bids ourselves. This gives us far more precision over which searches we're willing to pay for. Manual keyword targeting comes in three match types:
Self-ASIN campaigns use product targeting rather than keywords we place ads directly on relevant product detail pages, including our own client's other listings. This can help strengthen a brand's presence across its own catalogue and put a product in front of shoppers who are already looking at something similar. We're careful not to overstate what this achieves: it's a visibility and defensive tool, not a guaranteed sales driver, and its usefulness depends heavily on the product category and catalogue size.
| Campaign Type | Main Purpose | How It Works | How We Use It |
|---|---|---|---|
| Auto | Discovery | Amazon matches the product to relevant searches and targets automatically | Surface new keyword and targeting ideas, especially early on |
| Manual | Control & precision | We choose specific keywords or products and set bids ourselves | Scale proven, converting search terms with tighter control |
| Self-ASIN / Product Targeting | Visibility & defence | Ads placed on specific product detail pages, including a brand's own listings | Strengthen presence on relevant detail pages and capture comparison shoppers |
Targeting, Explained Simply
Targeting is the set of rules that decides where and to whom an ad is shown. We treat the different targeting types as parts of one connected system, not isolated settings a keyword that performs well in an Auto campaign often becomes tomorrow's Manual Exact target, and a competitor ASIN we're monitoring today might become a defensive target next month.
| Targeting Type | What It Does |
|---|---|
| Keyword targeting | Bids on specific search terms a shopper types into Amazon |
| Broad match | Widest reach; matches related and synonymous searches |
| Phrase match | Matches searches containing a set phrase, with some flexibility around it |
| Exact match | Matches only the specified search term (and very close variants) |
| Negative keywords | Excludes specific search terms from triggering an ad, to cut wasted spend |
| Product / ASIN targeting | Places ads on specific product pages rather than search terms |
| Category targeting | Targets an entire product category, useful for broader visibility |
| Auto targeting | Amazon selects relevant searches and products automatically |
| Competitor product targeting | Places ads on competitor product detail pages |
| Own product / Self-ASIN targeting | Places ads on the brand's own related listings |
How We Use ACOS
ACOS (Advertising Cost of Sale) is one of the main numbers we watch, but it's rarely the only one. It measures how much was spent on advertising compared to the sales that advertising generated:
Example: €100 spent on ads generating €500 in attributed sales →
€100 ÷ €500 × 100 = 20% ACOS
A lower ACOS generally means advertising is more cost-efficient but "low" isn't automatically "good," and we don't treat the lowest possible ACOS as the goal in itself. The right ACOS target depends on the product's margin, its price point, category competition, and whether the client's priority is profitability, growth, or launching a new product where some short-term inefficiency is expected. A 25% ACOS can be perfectly healthy for a high-margin product chasing growth, while a 10% ACOS might still be too high for a thin-margin item.
We also make sure clients understand how ACOS relates to the other numbers around it:
- Traffic - the impressions and clicks an ad generates; visibility, not yet a result.
- Advertising sales - revenue directly attributed to ad clicks.
- Orders - the number of purchases that resulted from those clicks.
- Conversion - the percentage of clicks that turn into a purchase.
- ACOS - the cost of generating that advertising sales figure.
High traffic with low conversion usually points to a targeting or listing problem, not a budget problem. That distinction is often the difference between "spend more" and "fix the targeting" as the right next step.
ACOS - one account, one performance window
Ad spend vs. advertising sales
One-Month Performance Snapshot
Here's a real example of what this looks like in practice, taken from Amazon Ads Campaign Manager for one account we manage. The dashboard shows two different date ranges, which we've kept clearly separated below - combining them would be misleading, since they cover different periods of different lengths.
Overview 1 Aug to 31 Aug 2026 (full month, all Sponsored Ads)These are the overview totals for the full month across all Sponsored Ads. "New-to-brand" purchases specifically count orders from customers who hadn't bought from this brand on Amazon in the past 12 months a useful signal for how much advertising is expanding the customer base, not just re-selling to existing buyers.
Reach narrowing into purchases (1–31 Aug 2026)
These numbers describe advertising performance, not overall profitability. Whether this level of ACOS and sales is "good" for the business also depends on product margin, cost of goods, and other expenses outside of Amazon Ads figures we don't have visibility into from the ads dashboard alone.
Reading the performance graph
The chart above tracks four lines over roughly three weeks: total cost, ACOS, sales, and purchases. It's normal for these to move independently rather than in a straight line a spike in cost on one day doesn't always produce a matching spike in sales that same day, since attribution can lag slightly and daily demand naturally fluctuates. What we're watching for over time is the overall trend: is ACOS drifting up or down, and is that being driven by rising cost, changing sales, or both. A single day rarely tells the full story; the trend across the window does.
Managing Multiple Amazon Marketplaces
We manage Amazon advertising across 15 marketplaces:
A campaign strategy that works well in one marketplace rarely transfers unchanged to another. Search volume, competition, cost-per-click, and language all vary by country, and so does customer behaviour a keyword that converts well in the US might be searched completely differently in Germany or Japan. Because of this, we build and optimise campaigns per marketplace rather than duplicating a single template everywhere, even when it's the same product being advertised.
Our Optimization Process
Once a campaign is live, the real work is ongoing. Here's the workflow we follow for most accounts:
Keyword and competitor research
Understand how customers search and what competitors are targeting.
Campaign structure
Build a structure that separates discovery (Auto) from control (Manual).
Launch
Go live with initial bids and budgets based on research.
Collect data
Gather search-term and performance reports over a set period.
Identify converting keywords / ASINs
Find what's actually driving efficient sales.
Add to controlled campaigns
Move proven targets into Manual campaigns with tighter bid control.
Add negatives
Exclude irrelevant or wasteful search terms and targets.
Adjust bids and budgets
Reallocate spend based on what the data supports.
Monitor ACOS and sales
Track performance against the account's actual goals.
Scale what works
Increase budget behind proven, efficient campaigns.
Common Amazon Ads Mistakes
Most underperforming accounts we take over share a handful of recurring issues:
Final Takeaway
Successful Amazon advertising isn't one clever setting or a single winning keyword it's the combination of research, structure, targeting, data, and continuous testing, applied consistently over time. Accounts that treat Amazon Ads as "set once and leave alone" tend to plateau; the ones that keep refining based on what the data actually shows tend to keep improving.
Wondering how your own Amazon Ads account stacks up?
If you're unsure whether your current campaign structure, targeting, or ACOS is working as hard as it could be, we're happy to take a look and talk through what we see.
Get in touch with our Ads team